You are viewing a plain text version of this content. The canonical link for it is here.
Posted to yarn-issues@hadoop.apache.org by "Subru Krishnan (JIRA)" <ji...@apache.org> on 2017/05/30 20:59:04 UTC
[jira] [Comment Edited] (YARN-3666) Federation Intercepting and
propagating AM-RM communications (part one: home RM only)
[ https://issues.apache.org/jira/browse/YARN-3666?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16030001#comment-16030001 ]
Subru Krishnan edited comment on YARN-3666 at 5/30/17 8:58 PM:
---------------------------------------------------------------
Thanks [~botong] for the updated patch. It's pretty close, have a couple of minor comments:
* We have both a ConcurrentHashMap and synchronized access for _containerIdToSubClusterIdMap_, I feel the latter is redundant.
* In *cacheAllocatedContainers*, we need a more nuanced approach than simply throwing exception on encountering containerId clash. We have to think about either reverting all (or selective) cached containers and/or continue with the remaining containers in the same allocation. Since this is an extreme corner case, I am fine with dealing with it fully in a separate JIRA.
* I think you missed my comments on code documentation.
was (Author: subru):
Thanks [~botong] for the updated patch. It's pretty close, have a couple of minor comments:
* We have both a ConcurrentHashMap and synchronized access for _containerIdToSubClusterIdMap_, I feel the latter is redundant.
* I think you missed my comments on code documentation.
> Federation Intercepting and propagating AM-RM communications (part one: home RM only)
> -------------------------------------------------------------------------------------
>
> Key: YARN-3666
> URL: https://issues.apache.org/jira/browse/YARN-3666
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: nodemanager, resourcemanager
> Reporter: Kishore Chaliparambil
> Assignee: Botong Huang
> Attachments: YARN-3666-YARN-2915.v1.patch, YARN-3666-YARN-2915.v2.patch, YARN-3666-YARN-2915.v3.patch, YARN-3666-YARN-2915.v4.patch, YARN-3666-YARN-2915.v5.patch, YARN-3666-YARN-2915.v6.patch, YARN-3666-YARN-2915.v7.patch
>
>
> In order, to support transparent "spanning" of jobs across sub-clusters, all AM-RM communications are proxied (via YARN-2884).
> This JIRA tracks federation-specific mechanisms that decide how to "split/broadcast" requests to the RMs and "merge" answers to
> the AM.
> This the part one jira, which sets up the basic structure, without secondary subclusters. All requests are forwarded to home subcluster. Part two is in YARN-6511
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: yarn-issues-help@hadoop.apache.org